ABSTRACT
Endocytosis plays important roles in regulating EGFR signaling. We previously found that EGFR endocytosis during mitosis is mediated differently than during interphase. While the regulation of EGFR endocytosis in interphase is well understood, little is known regarding the regulation of EGFR endocytosis during mitosis. Here, we studied the mechanisms regulating mitotic EGFR endocytosis. We found that contrary to interphase cells, mitotic EGFR endocytosis is more reliant on the activation of the E3 ligase CBL. At high EGF doses, inhibition of inhibited EGFR endocytosis of mitotic cells, but not of interphase cells. Moreover, the endocytosis of mutant EGFR Y1045F-YFP was strongly delayed. The endocytosis of truncated EGFR Δ1044-YFP that does not bind to CBL was completely inhibited. EGF induced stronger ubiquitination of mitotic EGFR than interphase EGFR and mitotic EGFR is trafficked to lysosome for degradation. Furthermore, during mitosis low doses of EGF also stimulate EGFR endocytosis by NCE. Contrary to interphase, CBL and the CBL-binding regions of EGFR were required for mitotic EGFR endocytosis at low doses. This may be due to the mitotic ubiquitination of the EGFR even at low EGF doses. In conclusion, mitotic EGFR endocytosis solely proceed through CBL-mediated NCE.
INTRODUCTION
The epidermal growth factor (EGF) receptor (EGFR), like other receptor tyrosine kinases (RTKs), regulates key events in cell growth, differentiation, survival and migration [1–3]. Aberrant signaling from EGFR has been implicated in many diseases [2, 4]. EGFR is historically the prototypical RTK. It was the first of this large family of transmembrane receptors to be cloned, and the first for which a clear connection between aberrant receptor function and cancer could be drawn [2]. The binding of EGF to EGFR at the cell surface induces dimerization of EGFR, which results in the activation of EGFR tyrosine kinase and EGFR trans-autophosphorylation [5, 6]. EGFR activation stimulates various signaling pathways that regulate multiple cell functions [1, 3]. EGF stimulates cell proliferation by driving the cell cycle that is comprised of four phases: G1, S, G2 and M [7, 8]. Binding of EGF also stimulates the rapid internalization of EGFR [9]. EGFR endocytosis and EGFR-mediated cell signaling are mutually regulated [10, 11].
In spite of significant advances in our understanding of EGFR signaling and trafficking, some critical knowledge is still lacking. Our current knowledge of EGFR signaling and EGFR endocytosis comes mostly from the studies of cells in G1 phase of the cell cycle. Very little is known regarding EGFR-mediated signaling and endocytosis in mitosis.
Mitosis represents a period where the needs and requirements of the cell differ vastly from interphase cells. EGFR signaling has been shown to be regulated differently between interphase and mitotic cells. We and others previously found that the EGFR of mitotic cells can still be activated during mitosis, but that the signal transduction pathways are regulated differently compared to interphase cells [12, 13]. We also previously found that EGFR endocytosis of mitotic cells is regulated differently, in that EGFR is endocytosed at a slower rate [14]. At the time, we did not fully decipher the molecular mechanisms behind the differential kinetics. Therefore, in this report, we further studied this phenomenon.
Endocytosis of the EGFR can lead to two distinct fates for the receptor: recycling back to the plasma membrane or lysosomal degradation. As such, the route taken directly influences the total number of receptors available for a subsequent signal transduction response. EGFR recycling has been shown to be mediated by clathrin-mediated endocytosis (CME), whereas non-clathrin mediated endocytosis (NCE) targets receptors for lysosomal degradation.
CME is a mechanism of internalization that is dependent on the recruitment of clathrin to the receptor. While this notion has been disputed by some studies [15–17], most data support the theory that CME is inhibited in mitosis [18–27]. The mechanisms underlying CME inhibition are still unknown, however, several mechanisms have been proposed and partially tested. These mechanisms include “moonlighting” hypothesis [25, 26], the phosphorylation of endocytic proteins [28, 29], and the unavailability of actin for CME [30]. In agreement with this, we previously found that mitotic EGFR endocytosis was clathrin-independent as siRNA depletion of clathrin heavy chain did not affect mitotic EGFR endocytosis [14]. We therefore hypothesized that mitotic EGFR proceeded exclusively through NCE.
NCE has been described as having potential tumor suppressive characteristics {Caldieri, 2017}. NCE has also been described as initiating more slowly than CME [9, 31–34], which fits with our observed delay in mitotic EGFR endocytosis [14]. Molecularly, EGFR NCE has been described as only activated by physiologically high doses of EGF [32, 35, 36], which is likely a mechanism evolved to compensate when the CME pathway is saturated and to prevent excessive EGFR signaling [37]. EGFR NCE has been shown to be mediated by ubiquitination of the receptor, and this ubiquitination has been shown to be limited by the activity of the E3 ligase c-CBL [32, 35, 36]. Therefore, c-CBL (henceforth CBL) provides a critical negative regulatory control of the EGFR, as it targets the EGFR for endocytosis and degradation. The activation of CBL depends on its binding to the activated EGFR, either by direct interaction with pY1045, or by indirect interaction through the adaptor GRB2, which binds to pY1068 or pY1086 [35, 36, 38, 39].
In this report, we find that EGF-stimulated EGFR endocytosis proceeds exclusively by NCE during mitosis. We find that contrary to interphase cells, mitotic EGFR endocytosis is more reliant on the activation of CBL. At high EGF doses, inhibition of CBL by siRNA or mutation inhibited EGFR endocytosis of mitotic cells, but not of interphase cells. Moreover, the endocytosis of truncated EGFR Δ1044-YFP, which does not bind to CBL, was completely inhibited. EGF induced stronger ubiquitination of mitotic EGFR than interphase EGFR and mitotic EGFR is only trafficked to lysosomes for degradation. Furthermore, we found that during mitosis, low doses of EGF also stimulate EGFR endocytosis by NCE. Contrary to interphase, CBL and the CBL-binding regions of EGFR were required for mitotic EGFR endocytosis at low doses. This was due to the mitotic ubiquitination of the EGFR even at low EGF doses.
MATERIALS AND METHODS
Antibodies and chemicals
Antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA), including: mouse anti-EGFR (sc-373746), anti-pY99 (sc-7020), anti-CBL (sc-170), anti-Ubiquitin (sc-8017), anti-Cyclin B1 (sc-245), and anti-β-Tubulin (sc-5274), rabbit anti-GRB2 (sc-8034) and anti-SHC (sc-967), and goat anti pY1068 (sc-16804) and pY1086 (sc-16804). LAMP-2 (AF6228) antibody was from R&D Biosystems (Minneapolis, MN). The horseradish peroxidase (HRP)-conjugated secondary antibodies were from Bio-Rad (Hercules, CA) and the fluorescence-conjugated secondary antibodies were from Jackson ImmunoResearch (West Grove, PA). Goat anti-mouse immunoglobulin G (IgG) conjugated with agarose were from Sigma (St. Louis, MO). EGF was from Upstate Biotechnology.
Plasmid construction
The EGFR-YFP, EGFR-Y1045F-YFP, EGFR-Δ991-YFP, and EGFR-Δ1044-YFP constructs were described previously [40]. The c-CBL-YFP and 70z-CBL-YFP constructs were generous gifts from the Sorkin Lab.
Cell Culture, transfection, and treatment
HeLa, 293T, and MCF-7 cells were growth at 37°C in Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum and antibiotic/antimycotic solution maintained at 5% CO2 atmosphere. For transfection, MCF-7 cells in 24-well plates were transfected using LipofectAMINE 2000 reagent (Invitrogen, Carlsbad, CA) as per the manufacturer’s protocol, and 293T cells in 24-well plates were transfected using calcium phosphate precipitation with BES (N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid) buffer. MCF-7 and 293T cells were chosen due to their low levels of endogenous EGFR. Small interfering RNA-mediated silencing transfections were done using CBL siRNA (sc-29241; Santa Cruz Biotechnology, Santa Cruz, Calif) in HeLa cells as per the manufacturer’s protocol.
Mitotic cells were collected by gentle mitotic shake-off as previously described [13]. Briefly, cells were arrested in prometaphase by treating cells with nocodazole (200 ng/mL) in serum-free media for 16h. The nocodazole-arrested cells were treated with EGF (2 ng/mL or 50 ng/mL) for 5, 30, and 45 min, or not treated with EGF (0 min). The EGF-containing media was then removed and serum-free media was added. Cells were placed on ice and dislodged by gently tapping the plates for 5 min. The mitotic cell-containing media was centrifuged at 1000 rpm for 5 min. The obtained mitotic cells were then lysed with cold Mammalian Protein Extraction Reagent (M-Per) (Thermo Fisher Scientific Inc, Rockford, IL USA) buffer in the presence of phosphatase and protease inhibitors including 100 mm NaF, 5 mM MgCl2, 0.5 mM Na3VO4, 0.02% NaN3, 0.1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride, 10 μg/ml aprotinin, and 1 μM pepstatin A. To collect lysates for interphase cells, cells were serum-starved for 16h. Cells were then treated with EGFR for 5, 30, and 45 min. To ensure consistency with the mitotic treatment, the cells were also tapped on ice for 5 min to remove mitotic cells and then left on ice for 5 min. The remaining interphase cells were collected by scraping on ice in cold M-Per in the presence of phosphatase and protease inhibitors. For both interphase and mitotic cells, after lysing, the samples were centrifuged at 21,000 ×g and the supernatant was collected for immunoblotting.
Immunoprecipitation and immunoblotting
Immunoprecipitation experiments were carried out as described previously [40]. Interphase or mitotic cells were lysed with immunoprecipitation buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 1% NP40, 0.1% sodium deoxycholate, 100 mM NaF, 0.5 mM Na3VO4, 0.02% NaN3, 0.1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride, 10 μg/mL aprotinin, and 1 μM pepstatin A) for 15 min at 4°C. Cell lysates were then centrifuged at 21,000 ×g. The supernatant, containing 1 mg of total protein, were incubated with 0.8 μg of mouse monoclonal anti-EGFR antibody A-10 (Santa Cruz) for 2 h at 4°C with gentle mixing by inversion. Goat anti-mouse IgG conjugated with agarose was added to each fraction and incubated for 2 h at 4°C with gentle mixing by inversion. Next, the agarose beads were centrifuged, washed three times with immunoprecipitation buffer, and 2× loading buffer was added. The samples were boiled for 5 min at 95°C and loaded for SDS-PAGE for subsequent immunoblotting.
Immunoblotting was performed as previously described [41]. Briefly, protein samples were separated by SDS-PAGE and were transferred to nitrocellulose. The membranes were blocked for non-specific binding, and incubated with primary antibody overnight. The membranes were then probed with HRP-conjugated secondary antibody followed by detection with enhanced chemiluminescence solution (Pierce Chemical, Rockford, IL) and light detection on Fuji Super RX Film (Tokyo, Japan).
Indirect immunofluorescence
Indirect immunofluorescence was performed as previously described [14]. Cells were grown on glass coverslips and serum-starved for 16 h. After treatment without or with nocodazole (200 ng/mL for 16h) and without or with EGF for various indicated times, the cells were fixed with ice cold methanol for 10 min. The cells were then permeabilized with 0.2% Triton X-100 for 10 min on ice. Next, cells were blocked with 1% BSA for 1 h on ice. Cells were then incubated with primary antibody overnight at 4°C. Primary antibody anti-CBL was used at 1:50, and anti-pEGFR-Y1086 and anti-EGFR were used at 1:200. Cells were then washed three times with PBS, and incubated with rhodamine- or FITC-labeled secondary antibody for 1 h at 4°C. Cells were then washed three times PBS, followed by nuclear staining with DAPI (4’6-diamidino-2-phenylindole) (300 nM). Finally, cells were washed three times and mounted. Images were taken with DeltaVision deconvolution microscopy (GE Healthcare Life Sciences, Buckinghamshire). Quantification of EGFR internalization was performed using ImageJ as previously described [14]. Briefly, the cells were visualized by differential interface contrast (DIC). For each image, a large polygon (VL) was drawn along the outer edge of the cell membrane to represent the entire area of the cell. In addition, a small polygon (VS) was drawn along the inner edge of the cell membrane to represent the cell interior. The VL and VS values were calculated for either stains of EGFR, pEGFR, or for YFP (for EGFR-YFP mutants), and membrane EGFR percentage was obtained by the following equation:
RESULTS
CBL interaction with EGFR during mitosis
EGFR expression at the plasma membrane does not change from interphase to mitosis [13, 14, 42]. Previously, we found that similar to interphase, stimulation of nocodazole-arrested mitotic HeLa cells with high doses of EGF (50 ng/mL) induced the phosphorylation of the EGFR at all major tyrosine residues, including Y992, Y1045, Y1068, Y1086, and Y1173 [13]. Moreover, this also phosphorylated CBL to similar levels [13].
To confirm mitotic CBL activation by EGF stimulation, we observed CBL localization in mitotic HeLa cells by immunofluorescence microscopy. Immunofluorescence co-staining using anti-EGFR and anti-CBL antibodies revealed that CBL co-localizes with EGFR upon 5 min of 50 ng/mL EGF treatment in both interphase and mitotic cells (Figure 1A). Furthermore, immunoprecipitation of EGFR using a monoclonal anti-EGFR antibody of both interphase and mitotic cell lysates showed that mitotic cells stimulated with EGF for 5 mins had higher IPs of CBL with EGFR than interphase cells (Figure 1B). Interestingly, CBL IP with EGFR decreased after 30 min EGF in mitotic cells, whereas it increased for interphase cells, and continued increasing at 45 mins EGF. Most surprisingly however, was that ubiquitination of the EGFR was enhanced at all time points studied during mitosis compared to interphase (Figure 1B). Since CBL also binds EGFR indirectly through the EGFR adaptor GRB2, we also immunoblotted EGFR immunoprecipitates for GRB2 and SHC. The results showed that during mitosis, GRB2 and SHC also bind to EGFR following EGF stimulation (Figure 1B).
In summary, double indirect immunofluorescence revealed that both EGFR and CBL co-localize after EGF stimulation during mitosis. Co-IP experiments also showed that EGF stimulated the interaction between EGFR and CBL. In addition, EGFR is more strongly ubiquitinated by EGF stimulation during mitosis.
Effects of altering CBL activity during mitosis
CME has been shown to be inhibited during mitosis [27, 29, 30]. Therefore, we sought to discover whether altering CBL activity, the major mediator of NCE, would inhibit EGFR endocytosis. We first silenced CBL in HeLa cells by siRNA transfection and found that transfected mitotic cells had much less EGFR endocytosis following EGF (50 ng/mL) stimulation, as observed by immunofluorescence (IF) staining of activated EGFR (Figure 2A).
In comparison, transfected interphase cells were little affected. Similarly, in MCF-7 cells transfected with EGFR-YFP knockdown CBL by siRNA also inhibited mitotic endocytosis of EGFR (Figure 2B). To further verify the role of CBL, we used the dominant-negative 70Z-CBL-YFP mutant, which has a deletion of 17 amino acids that disrupts the RING finger structure making it unable to interact properly with ubiquitin-conjugating enzymes (E2 ligases) [43, 44]. The 70Z-CBL-YFP protein can still bind to the cytoplasmic tail of activated EGFR [43, 45–47]. Transfection with 70Z-CBL-YFP significantly inhibited EGF-induced EGFR endocytosis during mitosis, but not in interphase (Figure 3A). Quantification of the data from Figure 3A showed that mitotic cells transfected with 70Z-CBL-YFP retained EGFR at the plasma membrane when compared with non-transfected cells, even after 60 minutes of EGF treatment (50 ng/mL) (Figure 3B). Taken together, downregulating CBL activity decreased EGFR endocytosis in mitosis, but not in interphase. Therefore, CBL activity appears more important during mitotic EGFR endocytosis than during interphase.
We next sought to see whether CBL overexpression could increase the rate of endocytosis during mitosis. CBL overexpression in HeLa cells did not appear to induce endocytosis at earlier time points, nor increase the rate of EGFR internalization (Figure 3C). This is similar to interphase cells, where it was previously reported that overexpression of CBL did not increase the rate of EGFR internalization [46, 48].
Role of EGFR C-terminal domains for mitotic endocytosis
We previously showed that mitotic EGFR endocytosis requires EGFR kinase activity [14]. Treatment with the EGFR-tyrosine kinase antagonist AG1418 inhibited mitotic EGFR endocytosis, and washing away AG1478 restored endocytosis [14]. In contrast, interphase EGFR could still undergo endocytosis in the presence of AG1478 [14]. To further explore the role of EGFR kinase activity in the activation of CBL, we blocked EGFR activation with AG1478 for 1 h prior to EGF (50 ng/mL) treatment. As before, this treatment prevented EGFR endocytosis during mitosis as visualized by IF [14] (data not shown). We next examined CBL phosphorylation by immunoblotting with antibody to p-CBL and showed that treatment with AG1478 inhibited EGF-induced CBL tyrosine phosphorylation in mitotic cells. (Figure 4A). Therefore, EGFR kinase activity is required for CBL activation.
We next sought to investigate which EGFR domains were important for mitotic endocytosis. We made use of previously constructed YFP-tagged EGFR mutants including EGFR with Y1045F substitution (Y1045F-YFP, no direct CBL binding), EGFR truncated at 1045 (Δ1044-YFP, no CBL binding), EGFR truncated at 992 (Δ991-YFP, no internalization), and WT (EGFR-YFP) (Figure 5F). We transfected these constructs into MCF-7 or HEK 293T cells, since they express low amounts of endogenous EGFR, then observed the effects of EGF treatment on their plasma membrane localization using indirect immunofluorescence (Figure 4B-E & Figure 5A-E).
In non-EGF treated MCF-7 cells, all mutants exhibited high plasma membrane localization and low cytoplasmic localization during both interphase and mitosis (Figure 5A-E). For the cells in the interphase, treatment of EGF (50 ng/mL) for 30, 45, or 60 mins significantly increased the internalization of EGFR-YFP, Y1045F-YFP, and Δ1044-YFP, but not Δ991-YFP which is endocytosis deficient due to the lack of internalization motifs. The internalization levels of EGFR-YFP, Y1045F-YFP, and Δ1044-YFP during interphase at all three time points were all similar. In contrast, these mutants responded to EGF treatment differently from each other when cells were in mitosis. For the cells in mitosis, EGF stimulated strong endocytosis of EGFR-YFP, approximately two-thirds of EGFR-YFP was internalized following 30 min EGF treatment, with more EGFR-YFP becoming internalized at 45 and 60 min. The internalization of both Y1045F-YFP and Δ1044-YFP however were impaired in mitosis. In mitosis, no endocytosis of Y1045F-YFP mutants was observed following addition of EGF for 30 min, and only very low level of EGFR endocytosis at 45 min. Interestingly, a high proportion of them eventually became endocytosed at 60 min. However, no EGF-induced endocytosis of Δ1044-YFP mutants was observed even at 60 min following EGF addition (Figure 5). Similar results were observed when the experiments were repeated in 293T cells (Figure 4B-E).
Taken together, this data shows that the CBL-binding domains of the EGFR are more important for mitotic EGFR endocytosis than interphase. These results also suggest that GRB2 cooperation for indirect CBL-binding to EGFR dramatically increases mitotic EGFR endocytosis.
Endocytic trafficking of EGFR in mitosis
The endocytic pathway that the EGFR takes has been shown to influence the fate of the EGFR. CME has been shown to lead to EGFR recycling, whereas NCE targets receptors for lysosomal degradation [35, 36]. Since we observed that EGFR endocytosis during mitosis proceeds exclusively in CBL-mediated NCE, we hypothesized that EGFR endocytosis during mitosis should lead exclusively to lysosomal trafficking. To test this, we examined the colocalization of endocytic route markers with EGFR by fluorescence microscopy. The EGFR of both mitotic and interphase cells showed strong co-localization with EEA-1 and RAB5 after 30 min EGF treatment, indicating that the EGFR is trafficked to early endosomes (Figure 6). EEA-1 and RAB5 did not co-localize with any EGFR at the plasma membrane of either mitotic or interphase cells. Staining with the late endosomal markers LAMP-2 showed that EGFR and LAMP-2 colocalized after EGF stimulation for 60 min. These data indicated that EGFR was targeted to lysosomes through NCE during mitosis.
Low EGF doses activate mitotic EGFR NCE
The above experiments were all performed using high concentrations of EGF (50 ng/mL). During interphase, low doses of EGF only activates CME, whereas high doses activate both CME and NCE [32, 35, 36]. If this finding is also applied to mitosis, low doses of EGF should not induce EGFR endocytosis in mitosis as CME is inhibited in mitosis and only high doses of EGF activates NCE. However, we previously observed that low doses TR-EGF (2 ng/mL) could still lead to their internalization in mitotic HeLa and CHO cells in a similar pattern as high dose of EGF [14]. We therefore decided to determine if EGFR endocytosis at both high and low dose of EGF are regulated by CBL-mediated NCE.
We hypothesized that similar to high dose, low dose EGF is still able to stimulate EGFR endocytose through NCE in mitosis. To test this hypothesis, we first repeated experiments described in Figure 3-5 with low dose of EGF (2 ng/ml), and we indeed obtained similar results (Figure 6-7). As shown in Figure 6, in MCF-7 cells transfected with EGFR-YFP, EGF at low dose (2 ng/mL) induced EGFR endocytosis in both interphase and mitotic cells (Figure 7). The endocytosis of Y1045F-YFP was only observed at 60 min following EGF addition in mitosis, and Δ1044-YFP was deficient in endocytosis during mitosis (Figure 7). These data suggest that at low dose EGF, mitotic EGFR endocytosis require its interaction with CBL. To further determine the involvement of CBL in low dose EGF-stimulated EGFR endocytosis in mitosis, we transfected Hela cells with either wild type CBL-YFP or non-functional 70z-YFP (Figure 8). As in high dose EGF conditions, expression of 70z-YFP inhibited EGFR endocytosis in response to low dose EGF. However, expression of CBL-YFP did not affect EGFR endocytosis in mitosis in response to low dose EGF. Together these data suggest that EGFR endocytosis induced by low dose EGF is also mediated by CBL
We then examined why low dose EGF is able to stimulate EGFR endocytosis through CBL-mediated NCE in mitosis, but not in interphase. We examined whether this is due to CBL-mediated ubiquitination of EGFR. To this end, we treated Hela cells with EGF at 2 ng/mL or 50 ng/mL for 45 min and examined the ubiquitination of EGFR in both interphase and mitosis (Figure 9A). Co-immunoprecipitation of EGFR and immunoblotting for ubiquitin revealed that as before, high dose EGF (50 ng/mL) induced higher EGFR ubiquitination in mitosis more than in interphase. Low dose EGF (2 ng/mL) did not induce the ubiquitination of EGFR in interphase, as previously reported [32, 35, 36]. However, low dose EGF stimulation caused significant EGFR ubiquitination in mitotic cells. Moreover, the binding of CBL to EGFR followed the same pattern as ubiquitination, with CBL again binding to EGFR at low doses during mitosis, but not during interphase.
We also performed similar experiments with different times of EGF stimulation, at 5, 30, and 45 mins (Figure 9B). The phosphotyrosine-specific antibody pY99 was used to confirm EGFR phosphorylation. Blotting for ubiquitin revealed that during mitosis ubiquitination of EGFR occurred at 5 mins, and is sustained through to 45 mins in response to low dose EGF stimulation. In contrast, in interphase EGFR ubiquitination only occurred briefly at 5 min, ubiquitination was very weak at later time points. In addition, CBL and SHC are pulled-down with EGFR during both interphase and mitotic. Therefore, it appears that low dose EGF stimulation differentially ubiquitinates mitotic EGFR, and not interphase EGFR.
Since mitotic EGFR is strongly ubiquitinated at low doses of EGF, and ubiquitination has been associated with EGFR degradation, we hypothesized that low EGF doses could lead to EGFR degradation during mitosis. Total cell lysates of interphase and mitotic cells treated with low doses of and total EGFR levels were assayed by Western blot. Whereas interphase EGFR levels remain constant throughout 45 mins of low dose EGF treatment, we found that mitotic EGFR levels drop drastically with time (Figure 9C). Taken together, these results suggest that, unlike interphase, low doses of EGF activate CBL-mediated EGFR degradation in mitotic cells. Interestingly, by Western blotting, the CBL band appears smaller in mitotic samples than interphase samples.
DISCUSSION
Our results showed that EGF-induced EGFR endocytosis during mitosis proceeds exclusively by CBL-dependent NCE (Figure 10). NCE plays a major role in the regulation of EGFR fate by targeting it to lysosomes for degradation. Our research has uncovered a temporal period by which to exclusively target EGFR for degradation. This bypasses the receptor recycling pathway that is undesirable if the goal is EGFR attenuation, or if it is to deliver and keep a pharmacological agent into a cell [49]. Targeting mitotic cells is feasible for EGFR-overexpressing cancer cells, as these cells intrinsically undergo more cell proliferation. In addition, the population of mitotic cells can be increased by treatment with anti-mitotic drugs, such as the commonly used taxanes and vinca alkaloids. Therefore, mitotic cells of EGFR-overexpressing cells can be targeted more directly. Moreover, the FDA-approved EGFR antibody cetuximab has been shown to initiate receptor endocytosis [50]. Whether mitotic EGFR treated with EGFR antibodies are also internalized by NCE remains to be investigated. However, if it does, nano-conjugation of EGFR antibodies to pharmacological agents may provide a targeted approach to treating these cancers.
The study of EGFR NCE thus far has relied on the inhibition of clathrin, as well as the use of high doses of EGF to activate NCE. Our results indicate that mitotic EGFR endocytosis is exclusively through NCE. Thus, mitotic cells offer an alternate system for studying the NCE of the EGFR regardless of EGF dosage. In general, NCE is much more complicated and very little understood. It is no surprise that very little is known regarding the regulation of mitotic NCE of EGFR. Our findings that mitotic EGFR endocytosis is mediated by CBL through NCE at both high and low dose of EGF advanced our understanding on both EGFR endocytosis and NCE in general.
The theory that EGFR ubiquitination is absolutely necessary for endocytosis has been a subject of controversy [51–54]. Our research supports the notion that ubiquitination by CBL is important for NCE [32, 36]. Furthermore, our research provides strong support for the requirement of CBL and GRB2 binding to the EGFR in order to cause its ubiquitination [35, 36].
Our results argue that GRB2-mediated CBL binding is more important than direct CBL-binding during mitosis, as the internalization of Δ1044 mutant had significantly inhibited, whereas the internalization of Y1045F mutant had only slightly inhibited. Overexpression of CBL did not accelerate nor enhance mitotic EGFR endocytosis in response to EGF. This has also been recently demonstrated by in vivo experiments [55]. NCE has been reported to initiate more slowly than CME [9, 32–34, 43, 56], and it therefore appears that CBL overexpression is not the limiting factor to the speed of EGFR NCE. Other important mediators of NCE, for example EPS15, EPS15R, and EPSIN [32], or endoplasmic reticulum (ER)-resident protein reticulon 3 (RTN3) and CD147 [57] may warrant investigation. More importantly, the exact mechanism by which the ubiquitination of the EGFR induces internalization is still unknown, and studies to elucidate the precise molecular mechanism would be extremely impactful.
The inactivation of CBL has been shown to display pro-oncogenic features [58–62]. Moreover, common pro-oncogenic EGFR mutations L858R and L858R/T790M have impaired CBL-binding, slower endocytosis, and impaired degradation [63]. EGFRvIII, the most common variant in gliomas, also has a reduced interaction with CBL and thus impaired ubiquitination owing to hypophosphorylation of pY1045 [64, 65]. Here, we showed that CBL activity during mitosis is even more important, and its activity is enhanced compared to interphase cells. Evolutionarily, since mitotic cells do not have active CME [27, 29], the activation of NCE may have been even more critical during mitosis to suppress EGFR overactivation. A loss of CBL activity, whether by inactivating mutations to CBL or EGFR CBL-binding, would therefore have a more pronounced effect during mitosis, as the EGFR would continue to signal excessively. The functional role of mitotic EGFR activation is still not well known. It is unclear how abnormally sustained EGFR signaling during mitosis affects cellular processes, however, it does appear to help mitotic cancer cells resist nocodazole-mediated cell death [13].
We showed that the EGFR is more strongly ubiquitinated during mitosis at both low and high doses of EGF, suggesting that CBL activity is enhanced during mitosis. How can CBL be better primed to induce endocytosis, even at low concentrations of EGF during mitosis? It has been shown that CBL also acts as an adaptor in the CME pathway. Since CME is no longer active during mitosis, a possible explanation may be due to increased CBL protein availability, as the cellular pool of CBL no longer needs to divide its time between CME and NCE. Another explanation may be that CBL is modified during mitosis to be better primed for its E3 ligase activity. Interestingly, probing with the CBL antibody revealed that the mitotic CBL band appears smaller compared to interphase (Figure 8), although the exact significance is unknown. Another alternative possibility may revolve around the DUBs (deubiquitinating enzymes) that deubiquitinate EGFR. Fifteen DUBs have been reported to impact EGFR fate, although some may be deubiquitinating non-EGFR component, such as EPS15 [66]. These DUBs could be shut off during mitosis, causing ubiquitination to persist longer than during the interphase.
Mitosis represents a phase of tremendous transition to the cell. A significant change is the global phosphorylation of mitotic proteins by mitotic kinases. For example, studies have shown the mitotic phosphorylation of over 1000-6027 proteins, including 14,000-50,000 phosphorylation events depending on the study [67–69]. Interestingly, many phospho-sites overlap between EGF-stimulated cells and mitotic cells [69]. Indeed, various components of the EGFR signaling and endocytic pathways appear to play different roles in mitosis, a phenomenon known as moonlighting [28, 70]. This includes important members of EGFR CME, such as clathrin, dynamin, and AP-2 [28, 71–74]. Since these proteins and many others are moonlighting in mitosis-related processes, their availability to participate in EGFR endocytosis during mitosis may be compromised. This may also affect EGFR signaling. It has been shown that EGF-induced AKT activation requires EGFR residence in clathrin coated pits, but not internalization [75, 76]. We previously showed that only AKT2, and not AKT1, becomes activated following EGF stimulation during mitosis [13]. Since CME is shut down during mitosis, it can be speculated that the differential activation of AKT during mitosis is a consequence of the inability of clathrin to be involved in mitotic EGFR endocytosis. Therefore, the changes imparted by global mitotic phosphorylation and mitotic cell rounding cannot be discounted to EGFR signaling, and likely of other signaling receptors as well.
In our study, we made use of the microtubule depolymerizer nocodazole to arrest cells in mitosis. So far, nocodazole is still the most widely used drug for arresting cells in mitosis [69, 77–79]. We decided to use nocodazole in our research in order to obtain synchrony between our Western blots, co-IPs, and immunofluorescence experiments, as it has been shown that the substage of mitosis can influence the kinetics of endocytosis [14]. Previous research has showed that factors such as serum starvation, nocodazole, and other mitotic inhibitors could inhibit CME [80]. However, it should be noted that the researchers were evaluating transferrin receptors, which is endocytosed by constitutive endocytosis rather than the ligand-induced mechanism used by EGFR. Furthermore, our previous study that showed that clathrin downregulation by siRNA had no effect on mitotic EGFR endocytosis was performed without the use of nocodazole [14]. We have also previously shown that 16 h nocodazole treatment does not lead to significant cell apoptosis [13].
The EGFR uses various signaling pathways to achieve numerous pro-oncogenic cellular outcomes. Endocytosis downregulates EGFR signaling from the cell surface, but initiates intracellular signaling from endosomes [81]. In this way, endocytosis controls EGFR signaling, spatially and temporally, making it an indispensable part of receptor signaling. Therefore, the interplay between EGFR signaling and endocytosis critically determines cellular outcome.
In conclusion, our research further showed that mitotic EGFR endocytosis proceed through CBL-mediated NCE, which supports the notion that mitotic endocytosis is not completely inhibited, but proceeds through NCE. The unique property of mitotic EGFR endocytosis offers an important opportunity for developing cancer therapy that targets both EGFR and mitosis.
ACKNOWLEDGEMENTS
We thank Dr. A. Sorkin (Department of Cell Biology, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA 15261) for generously sending us the CBL constructs including c-CBL-YFP and 70z-CBL-YFP. This work was supported in part by grants from the Canadian Institutes of Health Research (CIHR). Ping Wee is supported in part by a scholarship from the Natural Sciences and Engineering Research Council.